Download Consequences of warming on tundra carbon balance determined by

LETTERS
PUBLISHED ONLINE: 16 MARCH 2014 | DOI: 10.1038/NCLIMATE2147
Consequences of warming on tundra carbon
balance determined by reindeer grazing history
Maria Väisänen1*, Henni Ylänne1,2, Elina Kaarlejärvi3, Sofie Sjögersten4, Johan Olofsson3, Neil Crout4
and Sari Stark1
Arctic tundra currently stores half of the global soil carbon (C)
stock1 . Climate warming in the Arctic may lead to accelerated
CO2 release through enhanced decomposition and turn Arctic
ecosystems from a net C sink into a net C source, if warming
enhances decomposition more than plant photosynthesis2 .
A large portion of the circumpolar Arctic is grazed by
reindeer/caribou, and grazing causes important vegetation
shifts in the long-term. Using a unique experimental set-up,
where areas experiencing more than 50 years of either light
(LG) or heavy (HG) grazing were warmed and/or fertilized, we
show that under ambient conditions areas under LG were a 70%
stronger C sink than HG areas. Although warming decreased
the C sink by 38% under LG, it had no effect under HG. Grazing
history will thus be an important determinant in the response
of ecosystem C balance to climate warming, which at present
is not taken into account in climate change models.
In northernmost Eurasia, the management of semi-domesticated
reindeer (Rangifer tarandus L., the same species as wild caribou
in North America) constitutes an important means of land-use,
and reindeer husbandry exerts a strong control over ecosystem
functioning3 . Grazing often increases the proportion of graminoids
at the expense of evergreen and deciduous shrubs, and may enhance
litter and soil C decomposition4,5 . Previous studies have shown that
reindeer grazing has the potential to mitigate the effects of climate
warming on vegetation by suppressing the growth of deciduous
shrubs that otherwise benefit from climate warming6,7 . Short-term
grazer exclosure and warming studies have also indicated that
Arctic grazers may limit the positive effect of warming on the gross
ecosystem production (GEP), and therefore, grazed areas under
warmer conditions are weaker C sinks (negative net ecosystem
exchange; NEE) than ungrazed areas8 . However, vegetation shifts
after herbivore exclusion are highly transient in time5,9 ; thus,
assessments of the impact of long-term grazing regimes on the
responses of ecosystem C balance to climate warming are needed.
The key questions are how the vegetation shifts induced by longterm grazing alter the effects of warming on GEP, and how grazing
interacts with the indirect mechanisms associated with warming,
such as the expansion of deciduous shrubs, that greatly affect
conditions for decomposition10 .
We examined the effects of long-term grazing history on
ecosystem responses to warming and fertilization in a subarctic
tundra heath in northern Norway (69◦ 310 N, 21◦ 190 E) during
2010–2012. We addressed two main research questions. First,
what is the effect of long-term grazing on C fluxes—NEE,
GEP and ecosystem respiration (Re ; mg CO2 –C m−2 h−1 )—under
ambient conditions? Second, how do responses of C fluxes to
warming and fertilization differ under different long-term grazing
intensities? To answer these questions, factorial warming and
fertilization experiments were established on two sides of a
reindeer pasture rotation fence built in the 1960s. Differing grazing
pressure over the past 50 years has formed two distinct types
of vegetation on the different sides of the fence; the vegetation
under light grazing (LG) is dominated by dwarf shrubs (Empetrum
nigrum ssp. hermaphroditum, Betula nana, Vaccinium vitis-idaea, V.
myrtillus and V. uliginosum), whereas the vegetation under heavy
grazing (HG) is dominated by graminoids(Carex spp., Deschampsia
flexuosa, Festuca spp. and Juncus trifidus) with rapid growth and
high litter decomposition rates11 . In 2010, we also established shortterm reindeer exclosures on HG tundra (HGexc) to test how a
sudden termination of grazing would influence C fluxes and interact
with warming and fertilization. We hypothesized the following: HG
tundra is a weaker sink for C (NEE less negative) via increased Re
because grazing enhances litter and soil C decomposition; warming
and fertilization increase the GEP to a greater extent under HG,
where graminoids respond rapidly to environmental change. NEE is
determined by the balance between GEP and Re , thus, we expected
the response of NEE to warming to be weaker under HG than LG;
and excluding reindeer grazing on HG will amplify the effects of
warming and fertilization on GEP, thus, we expected the response
of NEE to warming to be weakest under HGexc.
Consistent with our first hypothesis, the midday seasonal
averages showed that LG was a 70% stronger C sink than HG
under ambient conditions (NEE more negative under LG; Fig. 1a,
control plots; F2,21 = 4.99, P = 0.017). As predicted, NEE difference
between long-term grazing intensities was explained by higher
Re under HG than LG (Fig. 1b, control plots; F2,21 = 5.34, P =
0.013) rather than a difference in the GEP (Fig. 1c, control plots;
F2,21 = 0.23, P = 0.8). In contrast with long-term responses, the
short-term exclusion of reindeer had no detectable effects on
the tundra C balance (no difference between HGexc and HG;
Fig. 1a–c). The higher C sink under LG is in agreement with
studies showing a stronger C sink in shrub- than graminoiddominated tundra12,13 and findings of increased C sink strength due
to herbivore exclusion8,14 . However, short-term exclosure studies
have linked the effects of geese14 and ungulates8 to shifts in
the GEP, whereas our study found that the effect of grazing on
NEE was driven by a shift in Re . Higher Re under HGexc and
HG is probably mediated by increased decomposition. Grazing
increases the proportion of graminoids in relation to evergreen and
deciduous shrubs and bryophytes, increases the abundance of litter
1 Arctic
Centre, University of Lapland, P.O. Box 122, FIN-96101 Rovaniemi, Finland, 2 Department of Biology, University of Oulu, FIN-90014 Oulu, Finland,
of Ecology and Environmental Science, University of Umeå, SE-901 87 Umeå, Sweden, 4 School of Biosciences, University of Nottingham,
Sutton Bonington Campus, Sutton Bonington, Leicestershire, LE12 5RD, UK. *e-mail: [email protected]
3 Department
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange
© 2014 Macmillan Publishers Limited. All rights reserved.
1
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2147
LETTERS
a
a
Net ecosystem exchange (NEE)
LG
HG
−15
Total abundance
HGexc
LG
600
HG
500
−30
a
−45
−60
b
a
a
a
a
a
400
a
a
a
a
a
CTL W F WF CTL W F WF CTL W F WF
b
Ecosystem respiration (Re)
HGexc
LG
mg CO2−C m−2 h−1
120
b
100
80
a
b
0
a
a
a
100
b
b
a
b
300
200
HG
b
b
Hits/100 pins
mg CO2−C m−2 h−1
0
HGexc
CTL W
60
40
20
F
WF
CTL W
120
WF
CTL W
F
WF
Evergreen shrubs
Graminoids
Bryophytes
Deciduous shrubs
Herbs
Lichens
b
0
F
Litter abundance
HGexc
LG
HG
CTL W F WF CTL W F WF CTL W F WF
c
Gross ecosystem production (GEP)
HGexc
LG
0
100
HG
80
Hits/100 pins
mg CO2−C m−2 h−1
−25
−50
−75
−100
−125
−150
−175
a
a
a
a
b
a
a
b
b
40
a
b
b
20
CTL W F WF CTL W F WF CTL W F WF
Figure 1 | Midday (08:00–18:00) CO2 -C flux rates averaged over growing
seasons 2011–2012 in a subarctic tundra heath under light grazing (LG),
short-term reindeer exclusion (HGexc) and heavy grazing (HG) by the
reindeer. Differences in NEE (a), Re (b) and GEP (c) in control (CTL),
warmed (W), fertilized (F), and warmed and fertilized (WF) plots. Letters a
and b denote significant differences among treatments within grazing
intensities. LG was a stronger C sink than HGexc and HG, as a result of a
similar GEP, but higher Re under HGexc and HG than LG. Warming reduced
the C sink under LG (NEE less negative) but not under HGexc and HG,
because warming increased Re similarly in all grazing intensities
whereas it increased GEP only under HGexc and HG. Bars represent
mean ± s.e.m., n = 8.
(Fig. 2a,b and Supplementary Table 1), and enhances litter15 and
soil4 decomposition. Enhanced soil nitrogen16 and higher growing
season soil temperatures (9.0 ± 0.3 and 7.5 ± 0.3 ◦ C for HG and
LG, respectively; Supplementary Table 2) may further intensify Re
under HG. Thus, long-term grazing simultaneously alters a number
of ecosystem properties that have the potential to influence Re . These
ongoing grazer-mediated processes cannot be detected after shortterm experimental herbivore exclusion. As the long-term and shortterm effects of herbivory derive from different mechanisms and
processes, the long-term effects of herbivory are both quantitatively
and qualitatively different from the short-term effects.
2
60
0
CTL W
F
WF
CTL W
F
WF
CTL W
F
WF
Figure 2 | The total abundance of vegetation and litter in a subarctic
tundra heath under light grazing (LG), short-term reindeer exclusion
(HGexc) and heavy grazing (HG) by the reindeer. Average abundances of
plant functional groups (a) and litter (b) on the basis of point-intercept
analysis in control (CTL), warmed (W), fertilized (F), and warmed and
fertilized (WF) plots in August 2012 after three years of treatments.
Evergreen and deciduous shrubs were more abundant under LG, whereas
graminoids were more abundant under HGexc and HG. Warming increased
plant abundance in all grazing intensities. The abundance of litter was
higher under HGexc and HG than LG and significantly increased by
fertilization (Supplementary Table 1). Bars represent the mean for plant
abundance and mean ± s.e.m. for litter abundance, n = 8.
Consistent with our second hypothesis, warming increased GEP
under HGexc and HG, and had no effect under LG (Fig. 1c;
significant three-way G × W × F interaction; Supplementary
Table 3). The responses in Re to warming were consistent at all
grazing intensities (Fig. 1b and Supplementary Table 3), which in
combination with the varying effects on the GEP had a major effect
on the ecosystem C sink strength: warming under LG decreased the
C sink by 38% (NEE less negative), whereas the NEE under HGexc
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2147
Ecosystem respiration, 2011
HGexc
a
LG
mg CO2−C m−2 h−1
300
LETTERS
HG
200
100
0
−100
13/09
30/08
14/08
02/08
20/07
07/07
21/06
09/06
01/06
13/09
30/08
14/08
02/08
20/07
07/07
21/06
09/06
01/06
13/09
30/08
14/08
02/08
20/07
07/07
21/06
09/06
01/06
Date (dd/mm)
b
Ecosystem respiration, 2012
HGexc
LG
300
HG
mg CO2−C m−2 h−1
200
100
0
−100
10/09
08/08
17/07
28/06
19/06
05/06
10/09
08/08
17/07
WF
28/06
19/06
05/06
F
10/09
08/08
W
17/07
28/06
19/06
05/06
CTL
Date (dd/mm)
Figure 3 | Seasonal patterns in Re in a subarctic tundra heath under light grazing (LG), short-term reindeer exclusion (HGexc) and heavy grazing (HG)
by the reindeer. Rates were measured in control (CTL), warmed (W), fertilized (F) and warmed and fertilized (WF) plots simultaneously on different
grazing intensities in varying weather conditions on nine occasions during growing season 2011 (a) and on six occasions during growing season 2012 (b).
Warming increased Re in the early growing season to a stronger extent under LG than HGexc and HG (Supplementary Tables 4 and 5). Lines represent the
mean ± s.e.m., n = 8.
and HG remained unaffected (Fig. 1a; G × W × F interaction;
Supplementary Table 3). The warming effects on modelled seasonal
sums of NEE, GEP and Re agreed with the warming effects on
midday seasonal averages (Supplementary Fig. 1). This finding is
the first to demonstrate that the responses of the ecosystem C
balance to global climate warming may differ drastically depending
on the grazing history of the system. Another important finding was
that existing differences in the NEE between the grazing intensities
under ambient conditions were negated by warming; the NEE did
not differ between grazing intensities in the warmed plots (Fig. 1a
and Supplementary Table 3). The observed effects of grazing on
C balances under the present climatic conditions may thus not
be applicable in a warmer climate. Contrasting with our third
hypothesis, the short-term grazer exclusion had no interaction with
the effects of warming on C fluxes (Fig. 1a–c). Divergent short- and
long-term effects of grazing on the responses of C fluxes to warming
is a novel finding and shows that the short-term grazer exclusion
is not necessarily adequate for evaluating the long-term ecosystem
feedbacks due to grazing with respect to a changing climate.
In disagreement with our second hypothesis, experimental
fertilization had weak effects on the ecosystem C fluxes at all levels
of grazing intensity (Fig. 1 and Supplementary Table 3). As found
in previous studies17,18 , fertilization exerted a modest increase in Re ,
although this seemed to occur only on HG (c. 16%; Fig. 1b and
Supplementary Table 3). Although the main effect of fertilization
was weak, fertilization showed interactive effects on GEP and NEE
with warming and grazing intensity that may provide insights into
the mechanisms by which long-term grazing directs the effects of
warming on C fluxes. Under LG, warming or fertilization alone had
no effect on the GEP, whereas combined warming and fertilization
increased it by 22% (Fig. 1c; significant G × W × F interaction;
Supplementary Table 3). As warming increased the GEP only after
nitrogen (N) limitation was alleviated by fertilization, we suggest
that GEP under LG was co-limited by low soil N availability and
temperature. Increased GEP by combined warming and fertilization
also negated the effect of warming on NEE (Fig. 1a). The lack of
similar interactive effects of warming and fertilization under HGexc
and HG indicates that GEP under high grazing intensity may not
by limited by nutrients, probably as a result of the positive effect of
long-term grazing on soil N availability16 . These findings indicate
that the indirect effect of herbivory on the soil nutrient cycling
constitutes an important mechanism by which herbivores influence
the ecosystem’s response to warming.
Temporal patterns in the ecosystem C fluxes revealed a variation
in the strength of the warming effects throughout the growing
season; notably, the warming-induced increase in Re was most
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange
© 2014 Macmillan Publishers Limited. All rights reserved.
3
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2147
HG
0.5
0.4
0.3
0.2
0.1
20/08
28/07
13/07
17/06
20/08
28/07
13/07
17/06
20/08
28/07
13/07
4
HGexc
LG
0.6
17/06
pronounced during the early growing season (Fig. 3a,b and
Supplementary Tables 4 and 5), when the ecosystem also was a C
source during midday (NEE positive; Supplementary Fig. 2).
Although this did not affect the Re averaged over the entire
growing season, the early growing season responses of Re to
warming were significantly stronger under LG than HGexc and
HG (significant G × W interactions; Supplementary Table 5).
We propose that increased temperatures during the early growing
season may enhance the maintenance and growth respiration of
dwarf shrubs12 and strongly affect Re . As winter-time CO2 fluxes
are substantial in the annual C budget in northern ecosystems13 , it
was important to consider how grazing or experimental treatments
influenced CO2 release during the winter. We found no effects
of grazing or treatments on the winter-time soil respiration in
2010–2011 (Supplementary Table 3; 0.07 ± 0.001 g CO2 -C m−2
averaged over all grazing intensities and treatments), indicating that
the effects of grazing and warming were manifested only during the
growing season.
Experimental warming increased the total plant biomass
(Fig. 2a and Supplementary Table 1) and normalized difference
vegetation index (NDVI; Fig. 4 and Supplementary Table 6) at
all grazing intensities. The fact that warming similarly increased
the aboveground plant biomass at all grazing intensities while
increasing the GEP only under HGexc and HG most probably
results from a high resource allocation to roots rather than shoot
biomass5,19 . The plant biomass was higher, although not statistically
significantly (Supplementary Table 1), in the short-term grazer
exclosures (Fig. 2a), which agrees with the theoretical concept that
herbivory may limit plant biomass in tundra20 . Supporting earlier
findings that grazing drastically suppresses the warming-induced
increase in evergreen dwarf shrubs6,7 , warming under HGexc more
than doubled the abundance of evergreen dwarf shrubs while
exerting weaker effects under HG (Fig. 2a and Supplementary
Table 1). Fertilization under HGexc caused a strong increase
in graminoids (Fig. 2a and Supplementary Table 1), which get
suppressed by grazers under HG owing to the high palatability
of graminoids for reindeer21 . The responses of vegetation to the
short-term exclusion highlight that a sudden termination of heavy
grazing pressure at nutrient-rich tundra may cause rapid changes
in the aboveground vegetation22 .
Our study demonstrated for the first time that grazing history
(at a timeframe of 50 years) may largely regulate the responses of
the ecosystem balance to climate change. Given that most of the
tundra biome is grazed by reindeer or caribou, our finding alters
predictions regarding how a warmer climate will influence global
C cycles and feedbacks in response to the future climate. Reindeer
husbandry in northern Eurasia is an important traditional and
contemporary means of land-use, with reindeer densities averaging
between 1.2–1.7 reindeer km−2 (ref. 3). However, as grazing intensity
at decadal and centennial timescales has varied greatly at different
locations as a result of reindeer management practices or historical
events influencing reindeer population sizes and migration routes23 ,
different tundra sites possess highly variable grazing histories that
result from the combination of natural abiotic and biotic factors,
cultural practices and political decisions. Studies on past reindeer
grazing patterns at regional and pan-Arctic scales are extremely
scarce24,25 , forming a large deficit in our ability to assess historical
grazing pressures. Although grazing pressure by wild caribou in
Northern America is lighter (approximately 0.25–1.3 caribou km−2 ),
it still exerts a major influence on vegetation21,26 and ecosystem
responses to warming6,7,27 . Our results show that it is critical to
know the grazing history before current tundra C balances and
their responses to climate warming can be understood. A holistic
approach that integrates the past grazing patterns and the future
prospects of grazing intensity is needed to predict the effects of
climate warming. Other major grazers of the Arctic (for example,
Normalized difference vegitation index
LETTERS
Date (dd/mm)
CTL
W
F
WF
Figure 4 | Normalized difference vegetation index in a subarctic tundra
heath under light grazing (LG), short-term reindeer exclusion (HGexc)
and heavy grazing (HG) by the reindeer. NVDI was assessed in control
(CTL), warmed (W), fertilized (F), and warmed and fertilized (WF) plots
analysed on four occasions in growing season 2012. Warming increased
NDVI in all grazing intensities (Supplementary Table 6). Lines present the
mean ± s.e.m., n = 8.
moose, muskoxen, lemming, geese) also need to be considered.
We propose that future research to further understanding of
the relationship between grazing intensity and C balance under
warming could potentially provide tools to counteract increases in
C release from the tundra, which are at present subject to rigorous
debate owing to their potential for climate feedback.
Methods
The site was located above the tree line (600–700 m asl) on a northern slope of
Raisduoddar Fjell (69◦ 310 N, 21◦ 190 E) in suboceanic northern Norway. A
pasture rotation fence built in the 1960s divides the area into a lightly grazed
migration range (LG) and a heavily grazed summer range (HG) experiencing
intensive grazing during reindeer migration in August. In 2010, eight blocks
traversing the LG and HG side of the fence were selected with similar topography
and altitude within blocks. Four plots on both LG and HG, were assigned to the
following treatments: control (CTL), fertilized (F), warmed (W) and warmed and
fertilized (WF). Warming was achieved using hexagonal open top chambers
(OTCs) made from 1.5 mm thick and 0.4 m high polycarbonate deployed in the
early growing season (16 June 2010, 31 May 2011, 5 June 2012) and removed just
before reindeer arrival (7 August 2010, 4 August 2011, 8 August 2012). OTCs
increase daytime air (1.2◦ –1.8 ◦ C) and surface temperature (1 ◦ C) corresponding
to predicted climate in year 205027 but also affect humidity and wind28 . Here,
OTCs increased surface temperatures by 0.9 ◦ C on LG and by 1.8 ◦ C on HG, with
no effect on soil temperature (Supplementary Information). Fertilization was
applied as ammonium nitrate (NH4 NO3 ) equivalent to 10 g N m−2 in 1 l of water
in the early growing season (16 June 2010, 9 June 2011, 28 June 2012). There is a
potential bias due to a single N application29 . Nitrogen dosage corresponds to the
predicted soil N increase after a 7 ◦ C increase in air temperature30 . Plots on HG
were divided into grazed (HG) and ungrazed (HGexc) subplots protected from
grazing by short-term exclosures (height 0.9 m, mesh 40 × 40 mm) established
before reindeer migration and removed after all reindeer had left the area to
avoid winter-time snow accumulation. Some browsing, trampling and fertilization
were detected on HGexc due to early arrival of some reindeer. The experimental
design did not include short-term exclosures on LG owing to very low grazing
pressure. Mean temperature and precipitation in 2010 and 2012 were close to the
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange
© 2014 Macmillan Publishers Limited. All rights reserved.
NATURE CLIMATE CHANGE DOI: 10.1038/NCLIMATE2147
LETTERS
long-term average (−0.5 ◦ C, 500 mm–700 mm), whereas in 2011 the average
temperature was 1.5◦ –2.0 ◦ C warmer and precipitation was 125% of the average.
Plant and litter abundance was recorded in mid-Aug 2010–2012. A modified
point intercept method was used in positioning 50 cm wide rows of ten vertical
pins at 10 cm distance from each other (eight rows on LG, four rows on HGexc
and HG). The total number of hits per species was recorded, normalized to hits
per 100 pins, and pooled into growth forms (evergreen dwarf shrubs, deciduous
shrubs, graminoids, herbs, bryophytes and lichens).
NEE (light measurement) and Re (dark measurement; chamber covered with
an opaque white hood) were measured during growing seasons 2011 (nine times)
and 2012 (six times) using custom-made cylinders (2 mm thick transparent
polycarbonate, diameter 30 cm, height 39 cm) placed on plots and sealed to the
ground with a plastic hem to create a closed system while avoiding damage to the
vegetation. Air inside the chamber was mixed with a battery-driven fan.
CO2 -fluxes (ppm), relative humidity (%) and temperature (◦ C) were recorded
using a Vaisala Carbon Dioxide Probe (GMP343) and Humidity and Temperature
Probe (HMP75) fitted to the cylinder top. Data was logged every 15 sec
for 5 min using a Vaisala Measurement Indicator (MI70). Cylinders were vented
between measurements to restore the CO2 concentration to a basal level.
GEP was obtained by subtracting Re from NEE. The mean seasonal daytime
fluxes (mg CO2 -C m−2 h−1 ) of NEE, GEP and Re were calculated. The
seasonal sum of C fluxes was assessed by a modelling approach
(Supplementary Information).
Treatment effects on the midday average of NEE, GEP and Re and vegetation
were analysed by a mixed model with grazing (G), warming (W) and fertilization
(F) as fixed factors, block nested with grazing as a random factor and year (Y) as
a repeated factor. A Least Significant Difference (LSD) test was conducted to
denote significant treatment effects within grazing intensities. To verify the main
effect of grazing on CO2 fluxes, analysis was separately conducted for control
plots. Sampling date (D) was used as a repeated factor for the NDVI and daily
CO2 flux rates. Month (M) was used as a repeated factor for the soil temperature.
The treatment effects on wintertime Re were analysed without a repeated factor.
In cases of significant interactions among factors, treatment effects were
separately analysed for D, M or grazing intensity. Akaike’s information criteria
and residual plots were used to assess the model fit and transformations were
made to meet the assumptions of a mixed model when necessary. The degrees of
freedom are obtained by the Satterthwaite approximation. Statistical analyses
were conducted using PASW Statistics 18.0 for Windows (PASW).
14. Sjögersten, S., van der Wal, R., Loonen, M. J. J. E. & Woodin, S. J. Recovery of
ecosystem carbon fluxes and storage from herbivory. Biogeochemistry 106,
357–370 (2011).
15. Olofsson, J. & Oksanen, L. Role of litter decomposition for the increased
primary production in areas heavily grazed by reindeer: A litterbag experiment.
Oikos 96, 507–515 (2002).
16. Olofsson, J., Kitti, H., Rautiainen, P., Stark, S. & Oksanen, L. Effects of summer
grazing by reindeer on composition of vegetation, productivity and nitrogen
cycling. Ecography 24, 13–24 (2001).
17. Volk, M. et al. Subalpine grassland carbon dioxide fluxes indicate substantial
carbon losses under increased nitrogen deposition, but not at elevated ozone
concentration. Glob. Change Biol. 17, 366–376 (2011).
18. Christiansen, C. T., Svendsen, S. H., Schmidt, N. M. & Michelsen, A. High
Arctic heath soil respiration and biogeochemical dynamics during summer and
autumn freeze-in-effects of long-term enhanced water and nutrient supply.
Glob. Change Biol. 18, 3224–3236 (2012).
19. Hafner, S. et al. Effect of grazing on carbon stocks and assimilate partitioning
in a Tibetan montane pasture revealed by 13 CO2 pulse labeling. Glob. Change
Biol. 18, 528–538 (2012).
20. Eskelinen, A. Herbivore and neighbour effects on tundra plants depend on
species identity, nutrient availability and local environmental conditions.
J. Ecol. 96, 155–165 (2008).
21. Post, E. S. & Klein, D. R. Relationships between graminoid growth form and
levels of grazing by caribou (Rangifer tarandus) in Alaska. Oecologia 107,
364–372 (1996).
22. Ravolainen, V. T. et al. Rapid, landscape scale responses in riparian tundra
vegetation to exclusion of small and large mammalian herbivores. Basic Appl.
Ecol. 12, 643–653 (2011).
23. Tommervik, H., Dunfjeld, S., Olsson, G. A. & Ostby Nilsen, M. Detection of
ancient reindeer pens, cultural remains and anthropogenic influenced
vegetation in Byrkije (Borgefjell) mountains, Fennoscandia. Landscape Urban
Plan. 98, 56–71 (2010).
24. Josefsson, T., Hörnberg, G. & Östlund, L. Long-term human impact and
vegetation changes in a boreal forest reserve: Implications for the use of
protected areas as ecological references. Ecosystems 12, 1017–1036 (2009).
25. Räsänen, S., Froyd, C. & Goslar, T. The impact of tourism and reindeer herding
on forest vegetation at Saariselkä, Finnish Lapland: A pollen analytical study of
a high-resolution peat profile. Holocene 17, 447–456 (2007).
26. Zamin, T. J. & Grogan, P. Caribou exclusion during a population low increases
deciduous and evergreen shrub species biomass and nitrogen pools in low
Arctic tundra. J. Ecol. 101, 671–683 (2013).
27. Elmendorf, S., Henry, G. H. R., Hollister, R. D. & Björk, R. G. Global
assessment of experimental climate warming on tundra vegetation:
Heterogeneity over space and time. Ecol. Lett. 15, 164–175 (2012).
28. Bokhorst, S. et al. Microclimate impacts of passive warming methods in
Antarctica: Implications for climate change studies. Polar Biol. 34,
1421–1435 (2011).
29. Phoenix, G. K. et al. Impacts of atmospheric nitrogen deposition: Responses of
multiple plant and soil parameters across contrasting ecosystems in long-term
field experiments. Glob. Change Biol. 18, 1197–1215 (2012).
30. Mack, M. C., Schuur, E. A. G., Bret-Harte, M. S., Shaver, G. R. & Chapin, F. S.
Ecosystem carbon storage in Arctic tundra reduced by long-term nutrient
fertilization. Nature 43, 440–443 (2004).
Received 22 October 2013; accepted 27 January 2014;
published online 16 March 2014
References
1. Tarnocai, C. et al. Soil organic carbon pools in the northern circumpolar
permafrost region. Glob. Biogeochem. Cycle 23, GB2023 (2009).
2. Davidson, E. A. & Janssens, I. A. Temperature sensitivity of soil carbon
decomposition and feedbacks to climate change. Nature 440,
165–173 (2006).
3. Forbes, B. C. & Kumpula, T. The ecological role and geography of reindeer
(Rangifer tarandus) in northern Eurasia. Geogr. Compass 3,
1356–1380 (2009).
4. Stark, S., Strömmer, R. & Tuomi, J. Reindeer grazing and soil microbial
processes in two suboceanic and two subcontinental tundra heaths. Oikos 97,
69–78 (2002).
5. Olofsson, J., Stark, S. & Oksanen, L. Reindeer influence on ecosystem processes
in the tundra. Oikos 105, 386–396 (2004).
6. Post, E. & Pedersen, C. Opposing plant community responses to warming with
and without herbivores. Proc. Natl Acad. Sci. USA 105, 12353–12358 (2008).
7. Olofsson, J. et al. Herbivores inhibit climate-driven shrub expansion on the
tundra. Glob. Change Biol. 15, 2681–2693 (2009).
8. Cahoon, S. M., Sullivan, P. F., Post, E. & Welker, J. M. Large herbivores limit
CO2 uptake and suppress carbon cycle responses to warming in West
Greenland. Glob. Change Biol. 18, 469–479 (2012).
9. Olofsson, J. Short-and long-term effects of changes in reindeer grazing pressure
on tundra heath vegetation. J. Ecol. 94, 431–440 (2006).
10. Myers-Smith, I.H. et al. Shrub expansion in tundra ecosystems: Dynamics,
impacts and research priorities. Environ. Res. Lett. 6, 045509 (2011).
11. Hobbie, S. E. Temperature and plant species control over litter decomposition
in Alaskan tundra. Ecol. Monogr. 66, 503–522 (1996).
12. Cahoon, S. M., Sullivan, P. F., Shaver, G. R., Welker, J. M. & Post, E. Interactions
among shrub cover and the soil microclimate may determine future Arctic
carbon budgets. Ecol. Lett. 15, 1415–1422 (2012).
13. Euskirchen, E., Bret-Harte, M. S., Scott, G. J., Edgar, C. & Shaver, G. R. Seasonal
patterns of carbon dioxide and water fluxes in three representative tundra
ecosystems in northern Alaska. Ecosphere 3, 1–19 (2012).
Acknowledgements
We thank A. Niva and S. Aakkonen for their valuable help with the field experiments. We
thank Su. Katves and Si. Katves for assisting with vegetation recording and J. Hyvönen for
helping with the statistical analysis. This study was funded by the Academy of Finland
(project numbers 218121 and 130507).
Author contributions
S.St. and M.V. initiated and managed the field experiment. J.O. and S.Sj contributed to
the planning of the experiment. M.V. and H.Y. carried out C flux measurements with
contributions from S.St. and S.Sj. E.K. and J.O. were responsible for plant and NDVI
analyses. N.C. was responsible for modelling C fluxes. M.V. carried out the statistical
analyses. M.V. and S.St. wrote the manuscript, to which all other authors contributed with
discussions and text.
Additional information
Supplementary information is available in the online version of the paper. Reprints and
permissions information is available online at www.nature.com/reprints.
Correspondence and requests for materials should be addressed to M.V.
Competing financial interests
The authors declare no competing financial interests.
NATURE CLIMATE CHANGE | ADVANCE ONLINE PUBLICATION | www.nature.com/natureclimatechange
© 2014 Macmillan Publishers Limited. All rights reserved.
5